A significant problem facing the computer industry is heat. The higher the temperature a component operates at, the more likely it is to fail. Also, high temperatures, while not causing catastrophic failures, can create data processing errors. Operation at high temperatures can cause power fluctuations that lead to these errors within a central processing unit (CPU) or on the motherboard anywhere that data management is handled. Despite efforts at reducing waste heat while increasing processing power, each new CPU and graphics processing unit (GPU) released on the market runs hotter than the last. Power supply and motherboard components required to provide power and handle signal processing also are producing more and more heat with every new generation.
The use of liquids in cooling systems to cool computer systems is known. One known method of cooling computer components employs a closed-loop, 2-phase system 10 as illustrated in FIG. 1. The 2-phase system 10 is employed to passively cool the north 12 and south 14 bridge chips. The vapor travels through a tube 16 to a cooling chamber 18, the vapor turns back into liquid, and the liquid is returned by tube 20 to the chips 12, 14 for further cooling. In another known liquid cooling system, internal pumps move liquid past a hot plate on a CPU and then the heated liquid is pumped into a finned tower that passively cools the liquid and returns it to the plate.
In the case of large-scale, fixed-installation supercomputers, it is known to submerge the active processing components of the supercomputer in inert, dielectric fluid. The fluid is typically allowed to flow through the active components and then it is pumped to external heat exchangers where the fluid is cooled before being returned to the main chamber.
Despite prior attempts to cool computer components, further improvements to cooling systems are necessary.